Systems and methods for wear assessment and part replacement timing optimization

Abstract

A method for part replacement timing optimization. The method includes training a wear estimate model. Training the model includes predicting a plurality of wear patterns for a part, each wear pattern corresponding to a degree of severity. Training images are rendered for each wear pattern. Each of the training images is labeled with the corresponding degree of severity. A neural network is then trained with the labeled training images. An image of a deployed part associated with a machine is received and fed into the trained wear estimate model. The method further includes receiving a wear estimate for the part image from the trained wear estimate model, estimating a change in performance of the machine based on the wear estimate, and determining a machine utilization pattern for the machine. The machine utilization pattern and the change in performance estimate are combined to determine an optimal time to replace the part.

What is claimed is: 
     
       1. A method for part replacement timing optimization, comprising:
 training a wear estimate model, including:
 predicting a plurality of wear patterns for a part, each wear pattern corresponding to a degree of severity;
 automatically generating, by executing a parametric modeling operation, a plurality of training images, each training image in the plurality of training images representing a combination of a corresponding one of the plurality of predicted wear patterns, the degree of severity, and at least one of a load type and a cycle time; and 
 
 labeling each of the plurality of training images with the corresponding degree of severity and the at least one of the load type and the cycle time; 
 training a neural network with the plurality of labeled training images; 
 receiving a part image of a deployed part associated with a machine; 
 feeding the part image into the trained wear estimate model; 
 receiving a wear estimate for the part image from the trained wear estimate model; 
 estimating a change in performance of the machine based on the wear estimate; 
 determining a machine utilization pattern for the machine; and 
 combining the machine utilization pattern and the change in performance estimate to determine an optimal time to replace the part. 
 
 
     
     
       2. The method of  claim 1 , wherein predicting the plurality of wear patterns comprises using a physics-based wear model. 
     
     
       3. The method of  claim 1 , wherein training the wear estimate model further comprises supplementing the plurality of training images with a plurality of labeled training photos of used parts. 
     
     
       4. The method of  claim 1 , wherein determining the optimal time to replace the deployed part comprises calculating a time period after which the cost of continuing to run the machine with the deployed part exceeds the total cost to replace the deployed part. 
     
     
       5. The method of  claim 1 , wherein determining the machine utilization pattern for the machine comprises receiving telemetry data from the machine. 
     
     
       6. The method of  claim 1 , further comprising identifying at least one of the machine or the deployed part. 
     
     
       7. A part replacement timing optimization system,
 comprising:
 one or more processors; and 
 
 one or more memory devices having stored thereon instructions that when executed by the one or more processors cause the one or more processors to: 
 train a wear estimate model, including operations to:
 predict a plurality of wear patterns for a part, each wear pattern corresponding to a degree of severity;
 automatically generate, by executing a parametric modeling operation, a plurality of training images, each training image in the plurality of training images representing a combination of a corresponding one of the plurality of predicted wear patterns, the degree of severity, and at least one of a load type and a cycle time; and 
 
 label each of the plurality of training images with the corresponding degree of severity and the at least one of the load type and the cycle time; 
 train a neural network with the plurality of labeled training images; 
 
 receive a part image, from an image capture device, of a deployed part associated with a machine; 
 feed the part image into the trained wear estimate model; 
 receive a wear estimate for the part image from the trained wear estimate model; 
 estimate a change in performance of the machine based on the wear estimate; 
 receive telemetry data from the machine; 
 determine a machine utilization pattern for the machine based on the telemetry data; and 
 combine the machine utilization pattern and the change in performance estimate to determine an optimal time to replace the part. 
 
     
     
       8. The system of  claim 7 , wherein predicting the plurality of wear patterns comprises using a physics-based wear model. 
     
     
       9. The system of  claim 7 , wherein training the wear estimate model further comprises supplementing the plurality of training images with a plurality of labeled training photos of used parts. 
     
     
       10. The system of  claim 7 , wherein determining the machine utilization pattern comprises applying a neural network to the telemetry data. 
     
     
       11. The system of  claim 7 , wherein determining the optimal time to replace the deployed part comprises calculating a time period after which the cost of continuing to run the machine with the deployed part exceeds the total cost to replace the deployed part. 
     
     
       12. The system of  claim 7 , further comprising instructions for identifying at least one of the machine or the deployed part. 
     
     
       13. One or more non-transitory computer-readable media storing computer-executable instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
 training a wear estimate model, including:
 predicting a plurality of wear patterns for a part, each wear pattern corresponding to a degree of severity; 
 rendering automatically generating, by executing a parametric modeling operation, a plurality of training images, each training image in the plurality of training images representing a combination of a corresponding one of the plurality of predicted wear patterns, the degree of severity, and at least one of a load type and a cycle time; and 
 labeling each of the plurality of training images with the corresponding degree of severity and the at least one of the load type and the cycle time; 
 training a neural network with the plurality of labeled training images; 
 
 receiving a part image of a deployed part associated with a machine; 
 feeding the part image into the trained wear estimate model; 
 receiving a wear estimate for the part image from the trained wear estimate model; 
 estimating a change in performance of the machine based on the wear estimate; 
 determining a machine utilization pattern for the machine; and 
 combining the machine utilization pattern and the change in performance estimate to determine an optimal time to replace the part. 
 
     
     
       14. The one or more non-transitory computer-readable media of  claim 13 , wherein predicting the plurality of wear patterns comprises using a physics-based wear model. 
     
     
       15. The one or more non-transitory computer-readable media of  claim 13 , wherein training the wear estimate model further comprises supplementing the plurality of training images with a plurality of labeled training photos of used parts. 
     
     
       16. The one or more non-transitory computer-readable media of  claim 13 , wherein determining the optimal time to replace the deployed part comprises calculating a time period after which the cost of continuing to run the machine with the deployed part exceeds the total cost to replace the deployed part. 
     
     
       17. The one or more non-transitory computer-readable media of  claim 13 , wherein determining the machine utilization pattern for the machine comprises receiving telemetry data from the machine. 
     
     
       18. The one or more non-transitory computer-readable media of  claim 17 , wherein determining the machine utilization pattern comprises applying a neural network to the telemetry data. 
     
     
       19. The one or more non-transitory computer-readable media of  claim 13 , further comprising identifying at least one of the machine or the deployed part. 
     
     
       20. The method of  claim 1 , comprising labeling each of the plurality of training images with the corresponding degree of severity, the load type and the cycle time.